Members of the University of Bolton Computer and Cyberpsychology Research Unit (UBCCRU) have recently developed The BARTA (University of Bolton Affect Recognition Tri-Stimulus Approach), a unique database providing validated images of facial affect. The BARTA’s unique use and combination of both synthetic and human faces to display facial expressions of emotion affords the academic community a new and comprehensive research resource that should facilitate new, innovative and pioneering research into the differential ways facial expressions of affect are perceived and processed when communicated by synthetic and human faces. The tri-stimulus feature also allows researchers to investigate the differential effects of synthetic emotion representation on atypical emotion perception, processing and recognition seen in many groups of clinical populations including those with autism spectrum disorders.

Comprising 395 validated digital colour images of emotional expressions, the BARTA is the first compilation to include three classes of face stimuli; emoticon, computer-generated-cartoon (CGC) and photographs of human faces. The BARTA includes multiple exemplars of the six basic emotions of happiness, sadness, anger, disgust, fear and surprise (Ekman & Friesen, 1977) portrayed by emoticon and CGC characters and actors of both genders and different ages and ethnicities. Over 200 adults and children assessed whether images accurately represent the intended affective state in a series of validation studies. Images were selected for inclusion in the BARTA if category membership was at 70% agreement (or above) amongst raters. The probability of inter-observer agreement was quantified with coefficient generalised kappa (p<.001).

 

The BARTA and autism

The database has already enabled a unique and systematic facial affect training package (“emotions-r-us”) to be developed by the BARTA research team specifically for children on the autistic spectrum. The development of the BARTA was guided by theory and findings on autism spectrum disorders (ASD) and therefore stimuli (including synthetic, cartoon faces) were designed to exploit the ASD individual’s social, aesthetic and learning preferences. The tri-stimulus approach (emoticon-cartoon-human) allows emotion recognition interventions to promote sequential learning by using stimuli types of increasing variance and complexity. Furthermore, synthetic faces may be the ideal medium to teach emotion recognition to children with autism, as behavioural and neuroimaging evidence suggest attentional, perceptual and processing atypicalities for human but not synthetic face stimuli in autism. It is argued by the BARTA research team that this evidence for social attention preferences for cartoon, but not human, faces in autism should drive the development of new emotion recognition interventions. Rather than presenting the to-be-learned facial expressions on human faces, as has been the case historically, interventions should exploit the known social preferences of autism by incorporating synthetic faces to depict emotional expressions. Presenting information on stimuli that is more socially relevant and which appears to be processed in a more typical and effective way, could increase motivation and exposure to facial expressions of emotion, allowing for enhanced appraisal and interpretation of those facial features used in the conveyance of emotional information and, as a result, may facilitate subsequent recognition and increase the chance of generalisation of this skill to human faces.

 

 

 

UBCCRU Home|

Cyberpsychology|

Individual Differences|

Technology Enhanced Learning|

UBCCRU Publications|

Cyberpsychology and computer psychology conference|